From a statistically complete sample of 188 radio galaxies in Abell clusters, we examine the radio/optical correlations, the FR I/II division, and the univariate and bivariate luminosity functions. As suggested by Owen (1993), the FR I/II division is shown to be a strong function of the optical luminosity of the host galaxy (∝ L 2 opt ). This dependence is also seen in the bivariate luminosity function, which suggests that the evolutionary tracks of radio sources and/or the initial conditions in the source are governed by the host galaxy properties. The probabilty for detecting radio emission increases with optical luminosity. The optical dependence is clearly separated in the integral luminosity functions which can be used as a constraint to models of FR I radio power evolution. Additionally, the source counts from the integrated univariate radio luminosity function (RLF) are consistent with our suggestion in paper V that radio sources may be a transient phenomenon which occurs in all elliptical galaxies at some time (or several times) over their lifetime. We find no statistically significant differences in the luminosity functions between rich cluster samples and radio sources not selected to reside in clusters. These results suggest that all radio galaxies live in similar environments in that the optical luminosity and the properties of the host galaxy are the most important parameters which affect radio source formation and evolution.

We present the results from a sensitive multiwavelength analysis of the properties of extremely red objects (EROs). Our analysis employs deep RIzJHK photometry of an 8<5 Â 8<5 region to select a sample of 68 EROs with ðRÀKÞ ! 5:3 and brighter than K ¼ 20:5 (5 ). We combine this photometric data set with an extremely deep 1.4 GHz radio map of the field obtained from the VLA. This map reaches a 1 limiting flux density of 3.5 lJy, making it the deepest 1.4 GHz map taken, and is sensitive enough to detect an active galaxy with L 1:4 e10 23 W Hz À1 at z > 1. If powered by a starburst, this radio luminosity is equivalent to a star formation rate of e25 M yr À1 for stars more massive than 5 M . We identify radio counterparts to 21 of the EROs in this field with radio fluxes above 12.6 lJy and resolve one-third of these with our 1>6 FWHM beam. The spectral energy distributions of the majority of these galaxies are consistent with those expected for dust-reddened starbursts at z $ 1. At these redshifts the radio luminosities of these galaxies indicate a median far-infrared luminosity of this population of L FIR e10 12 L , meaning that half of the radio-detected sample are ultraluminous infrared galaxies (ULIRGs). We conclude that e16% AE 5% of the ERO population brighter than K ¼ 20:5 are luminous infrared galaxies (LIRGs) at z $ 1. We also use photometric classification of the colors of the EROs to investigate the mix of dusty active and evolved passive systems in the remaining ERO population that is undetected in our radio map. Based on this we suggest that at least 30% and possibly up to $60% of all EROs with ðRÀKÞ ! 5:3 and K 20:5 are dusty, star-forming systems at ze1. Our best estimate of the star formation density in this highly obscured and optically faint (Re26) population is _ Ã ð0:1 100 M Þ ¼ 0:11 AE 0:03 M yr À1 Mpc À3 , comparable to estimates of that in H-emitting galaxies at z $ 1 and greater than the estimates from UV-selected samples at these epochs. This lends support to the claims of a strong increase in the contribution from obscured systems to the star formation density at high redshifts. Using the observed counts of the radio-detected ERO population, we model the apparent break in the K-band number counts of the whole ERO population at K $ 19 20 and propose that the passive ERO class dominates the total population in a relatively narrow magnitude range around Kd20, with dusty, active EROs making up the bulk of the population at fainter limits.